Solid-state and solution complexation of dopamine with the silicotungstic Keggin polyoxometalate

We investigated the interaction between the silicotungstic Keggin polyoxometalate (POM) and dopamine. Compound 1, [HDop 4 (SiW 12 O 40)]•4H 2 O, crystallized as intense deep-red crystals from an acidic aqueous solution containing silicotungstic acid (SiW) and 2 eq. dopamine after slow evaporation. Protonated dopamine (HDop) interacts with the POM via hydrogen bonding from both the ammonium hydrogens and the hydroxyl hydrogens. Dopamine itself forms “dimers” via π–π stacking between its aromatic rings. Size analysis by dynamic light scattering (DLS) of SiW with and without dopamine indicated strong interaction even in solution; this was supported by complementary diffusion-ordered (DOSY) nuclear magnetic resonance (NMR) spectroscopy. However, in solution, unlike in the solid-state structure, only one dopamine molecule closely associated with the POM. Temperature-dependent NMR revealed a strengthened association between silicotungstate and dopamine upon heating because of the weakening of hydrogen bonding between solvent water molecules.

Polyoxometalates (POMs) are discrete molecular oxo-anions (W, Mo, or V) or occasionally oxo-cations (Ga or Al) where the metal ion usually exhibits its highest oxidation states. A large number of POMs with different compositions and structures are known (Bijelic et al., 2018; Misra et al., 2019). Because of the very small sizes of POMs (about 1 nm) and the possibility of tailoring their chemistry (through grafting of ligands and changing of metal ions), there is currently a strong interest in the use of POMs in biological and biomedical applications (Lentink et al., 2023; Kong et al., 2020; Nefedova et al., 2024; Paredes-Perez et al., 2024). These include selective bond cleavage (Moussawi et al., 2025; Declerck et al., 2024), additives for protein crystallization (Bijelic and Rompel, 2018), and antibacterial agents (Gumerova et al., 2018). In one example, a transition-metal-substituted Wells–Dawson POM demonstrated a function as an inhibitor for the amyloid β (Aβ) aggregation which has been linked to the development of Alzheimer’s disease (Gao et al., 2014). Anderson–Evans POMs have shown tailorable interaction with phospholipid membranes governed by changes in addenda and central ions (Pashkovskaya et al., 2024). Decavanadate was found to function as an inhibitor for the cAMP-dependent protein kinase by binding its peptide substrate kemptide (Pluskey et al., 1997). In contrast, polyoxovanadates monosubstituted with Pt or Mo were found to stimulate the activation of the luteinizing hormone receptor (LHR, a G-protein-coupled receptor) by interaction with the cell membrane, promoting aggregation of LHR (Kostenkova et al., 2023).

While much attention has been focused on amino acids and peptides, POMs may interact with other biomolecules as well. One molecule of particular interest is dopamine. Dopamine plays an important role as a signal substance in the nervous system, acting as a neurotransmitter; it is prominently involved in the reward system but also affects mood and motivation. Chemically, dopamine is a catecholamine with a catechol base containing two vicinal hydroxyl groups that allow for strong complex formation. Dopamine has a good ability to coordinate strongly to metal oxide surfaces, e.g., iron oxide and titanium dioxide (Gu et al., 2015; Cadmen et al., 2022). Thus, due to their small sizes, metal oxide nanoparticles and POMs may interfere in molecular processes (Fadeel, 2019). For instance, dopamine could interact with POMs or metal oxides introduced for theranostic purposes or unintentionally via aerosol or dust exposure, which could interfere with different signaling pathways.

Quite a few studies on the application of various POMs as sensors for the detection of dopamine have been reported (Sun et al., 2018; Zhang et al., 2016), though studies on molecular interactions by single-crystal data are rarely reported. We have previously reported the interaction between the Keggin POM phosphotungstic acid (H 3 [PW 12 O 40)] and dopamine, which crystallized as [HPW 12 O 40 (HDop) 2 ]•4H 2 O (Svensson and Kessler, 2021). Furthermore, dopamine is known to promote the self-assembly of phosphotungstic acid into hierarchical 3D structures (Li et al., 2019; Svensson et al., 2021). A systematic study suggests that the ability to form hydrogen bonds at the opposite ends of dopamine, as well as π–π stacking due to aromatic rings, are important factors for dopamine’s function as a structure-directing agent (Svensson et al., 2021). To expand the knowledge about dopamine–POM interactions, we have employed polyoxometalate silicotungstic acid (H 4 [SiW 12 O 40 ]). Silicotungstic acid is isostructural to phosphotungstic acid, but the central ion has been replaced with silicon, altering the W–O bond polarity. We report herein the preparation and the solid-state and solution properties of the complex (HDop 4 [SiW 12 O 40 ]•4H 2 O).

Synthesis of compound 1

Silicotungstic acid (H 4 SiW 12 O 40 •H 2 O, 200 mg, 0.069 mmol) (Sigma-Aldrich, >95%) was dissolved in 5 mL 0.1 M HCl and mixed with a solution of two equivalents (eq., 26 mg, 0.138 mmol) of dopamine•HCl (Sigma-Aldrich, >99%) in 5 mL 0.1 M HCl. The final concentrations of silicotungstic acid and dopamine were 6.9 mM and 13.8 mM, respectively. The reaction mixture after mixing was transparent with a slight red tint and had a pH of 1.4. The solution was left in a plastic Petri dish to slowly evaporate. After about 3 weeks, massive formation of bipyramidal crystals of intense red color occurred. No precipitates were observed before crystallization. Replication of the synthesis afforded the same compound. The yield was about 20% based on the POM.

X-ray crystallography

Single-crystal data for crystals of 1 were collected using a Bruker D8 SMART APEX II CCD diffractometer (Billerica, USA) with λ(Mo − Kα) = 0.71073 Å and a graphite monochromator. The Apex4 program suite was used for data processing and refinement. The diffraction data were recorded at room temperature. Powder X-ray diffractogram was recorded for a crop of small ca. 0.1-mm intact crystals of 1 packed in a Lindeman capillary of 1 mm in diameter. Data were collected in the phi360 mode (scanning at 3° s –1). DIFFRAC.EVA v6.0.0.7 software was used for data processing and analysis. The theoretical XRD pattern for 1 was calculated from the cif file using Mercury version 2025-2.0 (CCDC).

Infrared spectrometry

A PerkinElmer FTIR spectrometer Spectrum 100 was used to record the IR spectrum. Washed and dried crystals of 1 were grinded to a fine red powder which was mixed with anhydrous KBr (dried at 200°C overnight) and pressed into a pellet. Spectra were averaged over 16 scans with a resolution of 2 cm –1. Reference spectra of silicotungstic acid hydrate and dopamine were recorded as well.

Scanning electron microscopy and energy-dispersive X-ray spectroscopy

For scanning electron microscopy, a Hitachi FlexSEM 1000 II microscope operating at an accelerating voltage of 5.00 kV and working distance of 5 mm was used. For energy-dispersive spectroscopy (EDS) analysis, an AZtecOneXplore EDS detector made by Oxford Instruments with an accelerating voltage of 15 kV and working distance of 10 mm was used. The sample crystal was mounted on carbon tape. EDS maps were collected over ca. 2 min.

Diffusion NMR and temperature series NMR

For nuclear magnetic resonance (NMR) analysis, a Bruker Avance III 600 MHz SmartProbe with TopSpin version 3.5 was used. Solution studies of silicotungstate with dopamine were done using freshly prepared mixtures, since compound 1 was not soluble in water or diluted hydrochloric acid. The NMR samples were prepared by dissolving 200 mg silicotungstic acid in 5 mL 0.1 M HCl containing 2 eq. of dopamine. Then, 450 µL of sample solution was transferred into a 3-mm-diameter NMR tube containing 50 µL D 2 O (Eurisotop). The spectra were processed and analyzed using TopSpin version 4.5.0. To investigate the stability of the silicotungstate–dopamine complex in solution, 1 H NMR spectra for the reaction solution were recorded at different temperatures (25 o C, 35 o C; 45 o C, 55 o C, 65 o C, and 70 o C). The diffusion-ordered (DOSY) NMR spectra were recorded at 25 °C.

Dynamic light scattering

A Malvern Zetasizer Nano ZS was used to estimate POM size in solution with and without the addition of dopamine. SiW or SiW with 2 equivalents of dopamine•HCl was dissolved in 0.1 M HCl. Data were recorded in triplicate at 25 o C. The concentrations of silicotungstic acid and dopamine in the solutions were 6.9 mM and 13.8 mM, respectively.

UV-visible spectrophotometry

UV-visible (UV-vis) spectra of silicotungstate (6.9 mM) with 2 eq. dopamine (13.8 mM) in 0.1 HCl were recorded at 25 o C and 70 o C in a Multiskan SkyHigh microplate spectrophotometer using a quartz cuvette. For the 70 o C measurement, the cuvette containing the analyte solution was heated in a stirred water bath to 70 o C, then immediately dried, and transferred to the spectrophotometer. The cuvette was closed during heating.

Crystallography

Reaction at room temperature between silicotungstic acid and 2 equivalents of dopamine in 0.1 M hydrochloric acid led to the formation of [HDop 4 [SiW 12 O 4 0]•4H 2 O (1) (Figure 1a) after about 3 weeks of slow evaporation in a Petri dish. Compound 1 crystallized in the orthorhombic space group P2(1)2(1)2(1). Detailed crystallographic data are presented in Table 1. The asymmetric unit consists of a silicotungstate anion surrounded by four dopamine molecules. The amine group of each dopamine is protonated, thus balancing the charge of the [SiW 12 O 40 ] 4− anion. Both catechol hydroxyl groups are protonated (i.e., in the –OH form). We previously reported the crystal structure of the Keggin phosphotungstate ion with dopamine (Figure 1b) (Svensson and Kessler, 2021). Phosphotungstic acid is isostructural to silicotungstic acid, with the main difference being the central atom. Differences in charge and electronegativity between P and Si contribute to the higher acidity of phosphotungstic acid; thermal programmed desorption of NH 3 adsorbed on silicotungstic acid and phosphotungstic acid has indicated stronger acidity in the latter (Kim et al., 2025).

Given the equal conditions for the preparation of the silicotungstic and phosphotungstic complexes with dopamine, the deep red color of the former would tentatively suggest a lower-energy charge-transfer transition compared with that of the phosphotungstate complex. This difference would originate from the change of the central atom, where the Si 4+ ion would promote a slightly larger polarity in the W–O bonds leading to a stronger charge-transfer interaction (Hood et al., 2025; Li et al., 2024).

Both in the case of the phosphotungstate complex with dopamine and in that of the silicotungstate complex with dopamine, all dopamine molecules coordinate in an edge-on mode; i.e., the aromatic ring is perpendicular to the POM. This supports the suggestion that the major interaction is via hydrogen bonding (as seen in the bond lengths below). A face-on coordination (the aromatic ring parallel to the POM) had instead indicated dominating π–metal interactions. Magueres et al. (2000) studied the formation of charge-transfer complexes between Lindqvist and Keggin POMs (including silicotungstate) with anthracenes and pyrenes containing a cationic ammonium or pyridinium anchor (i.e., bearing a structural resemblance to dopamine). In all cases, the organic ligands interacted with a POM oxygen via the cationic groups. Meanwhile, the aromatic parts of the molecules formed dimers via π–π stacking, which we also see both in compound 1 and in the previously reported [(HDop) 2 HPW 12 O 40 ] compound (Svensson and Kessler, 2021; Svensson et al., 2021). For compound 1, hydrogen bonds between the ammonium groups and bridging POM oxygens are found, (N2D)H2D2-O20 2.220 Å and (N3B)H3B2-O35 2.584 Å, as well as between a catechol hydroxyl group and bridging POM oxygen, (O3X)H3XA-O29 3.034 Å (Figures 1c–f).

[(HDop) 2 HPW 12 O 40 ], on the other hand, interacts with dopamine via the ammonium group and a terminal POM oxygen (O3) with a distance of 2.322 Å. One dopamine hydroxyl group hydrogen bonds a bridging POM oxygen (O17) with a bond length of 2.582 Å. The solvating water molecule O1E shows close contact with a dopamine hydroxyl group (2.447 Å) and another dopamine ammonium group (2.913 Å). The distance between the stacked dopamine aromatic rings is 3.256 Å. However, for [(HDop) 2 HPW 12 O 40 ], the stacking of the aromatic rings is quite offset compared with [(HDop) 2 HPW 12 O 40 ], where the rings are almost parallel. Supplementary Figure S1 highlights selected hydrogen bonding in the packed structure of 1. Additional crystallographic data for 1 are available in Supplementary Tables S1–S51.

Compound 1 was crystallized from a 0.1 M HCl solution, which obviously is far from physiological pH. The reason for this is easier crystallization than with near-physiological pH for this system (which formed powder precipitates instead of crystals at pH 7). However, due to the relatively high pK a values of the hydroxyl groups and the amino group (ranging from about 9 to 12 (Sanchez-Rivera et al., 2003)), dopamine will have the same cationic form at both pH 1 and pH 7. Hence the present conditions remain relevant for physiological conditions.

The silicotungstic POM is a suitable model for silica and silicate nanoparticles, which constitute the major component of sand and are common in everyday exposure via, for instance, aerosols (Davis et al., 1984). The silica (nano)particles have a negative charge at physiological pH, but they do not form complexes with catechols, unlike, for instance, TiO 2 and Fe 2 O 3. From the bond length O POM –N dopamine (3.0 Å), the bond energy was calculated to be 76.7 kJ/mol, which can be compared with the streptavidin–biotin non-covalent interaction commonly cited as about 100 kJ/mol (Ptaszek et al., 2025).

Powder XRD was recorded for small, intact crystals of 1 packed in a Lindeman capillary and compared with the simulated powder XRD pattern calculated from the cif file (Figure 2). The general trends of the experimental and theoretical diffractograms agree quite well. The experimental diffractogram, however, appears to have broadened compared with the theoretical diffractogram. This could be explained by a combination of factors, such as instrumental broadening and preferred orientation (crystal packing in the Lindeman tube). Grinding compound 1 into a fine powder impaired the crystallinity and produced a poorly crystalline diffractogram, presumably due to destruction and desolvation of the structure.

FTIR

The FTIR spectrum of a powdered sample of 1 was recorded along with references of dopamine and silicotungstic acid (Figure 3). A broad, strong band at ∼780 cm −1 is due to W–O–W bonds, and another rather strong band at ∼970 cm −1 is terminal W=O bonds. The contribution from the aromatic ring is seen at 1,505 cm −1 and 1,600 cm −1, attributed to C–H arom and C=C, respectively. The strong C=C band centered at 1,600 cm −1 presumably overlaps with the water signal commonly found at ∼1,620 cm −1. Between 3,100 cm −1 and 3,450 cm −1 is a complex region of multiple weak bands. This complex set of bands is likely a combination of multiple N–H bonds from the protonated amino group and hydrogen bonding between dopamine and the polyoxometalate (Fleming and Williams, 2007; Greijer et al., 2022). In the reference spectrum of dopamine, a distinct band is seen at 3,345 cm –1. This is due to a –NH 3 + group, as dopamine is in the form of a hydrochloride salt. In compound 1, the ammonium–chloride interaction changes to a –NH 3 +… O(W) hydrogen bond, causing an upshift of the absorption band.

SEM–EDS

SEM micrographs of compound 1 were recorded, as shown in Figure 4a, and compared with a digital photograph (Figure 4b), in which the pyramidal crystal morphology is clearly seen. EDS mapping was also performed (Figure 4c), confirming the presence of tungsten, silicon, nitrogen, and oxygen in the crystal. The crystal surfaces are relatively smooth and appear to have a layered structure.

Dynamic light scattering

Given the apparently strong interaction between silicotungstate and dopamine (as seen in the solid-state structure and from color change in the reaction mixture), it was interesting to see whether they associated even in solution. The experiment was carried out in the ratio of POM to dopamine of 1:2 that was used for crystallization of the material. This ratio was chosen because higher dopamine content results in the formation of opalescence and partial precipitation of the water-insoluble complex. Dynamic light scattering (DLS) was used to estimate the hydrodynamic diameter of the POM with and without the addition of dopamine. The diameter was found to have an average size of 1.04 nm (standard deviation 0.013 nm), which is in agreement with the literature value (Katryniok et al., 2010). When dopamine was added, the average diameter increased to 1.44 nm (standard deviation 0.11 nm). These results strongly indicate that dopamine is associated with silicotungstate even in solution, as is also hinted by the color change of the solution. The 0.3 nm (3 Å) increase in hydrodynamic size corresponds to the addition of a single dopamine molecule closely associated with the POM core (Figure 5).

NMR

As a complement to DLS, DOSY NMR was also used to study the interaction of compound 1 in solution. As 1 was not soluble in H 2 O/D 2 O, a mixture of SiW with 2 eq. of dopamine in 0.1 M HCl with 10% D 2 O was used instead, closely resembling the solvent conditions for preparing 1. A diffusion constant of 6.4E-10 m 2 /s was determined for silicotungstate with dopamine, about 32% smaller than for pure dopamine, 9.3E-10 m 2 /s. A smaller diffusion constant means slower movement in solution, which in this case is consistent with an increased size from the association between silicotungstate and dopamine in solution. Thus, this is in agreement with the DLS results.

A series of 1 H NMR spectra were recorded at increasing temperatures (25 o C, 35 o C, 45 o C, 55 o C, 65 o C, and 70 o C). A downshift in all hydrogen signals was noted, increasing with increasing temperature (Figure 6; Supplementary Figure S1). The signals were up-shifted about 0.1 ppm per 10° o C. At elevated temperatures, the hydrogen bond network in water weakens and allows dopamine to approach closer to the POM. This would mean an increased polarization in dopamine, causing the signal downshift. A concurrent broadening of the ammonium hydrogen signal at ca. 7.5 ppm reveals faster hydrogen exchange as temperature increases (Supplementary Figure S2). A reference spectrum for dopamine without silicotungstate is shown in Supplementary Figure S3. We hypothesized that a strengthened interaction between dopamine and silicotungstate would have an impact on the charge-transfer dynamics and could be probed by UV-vis spectroscopy. UV-vis spectra were recorded for the same reaction solution at 25° o C and 70 o C (Figure 7). A decrease in absorbance in the 400–500 nm range is seen at 70 o C. From the combined NMR and UV-vis spectra, we conjecture that upon heating, the water hydrogen-bond network weakens upon breaking of the hydration sphere around silicotungstate, allowing stronger interaction with dopamine via the ammonium group. Meanwhile, the increased temperature causes the aromatic ring to move more and/or change position relative to the POM, causing a decreased charge-transfer, which is reflected in decreased absorption of the complement color region.

The thermal energy of the molecules can be estimated by multiplying the general gas constant with the temperature in kelvin (RT) at ca. 2.5 kJ/mol and 2.9 kJ/mol for 25 o C and 70 o C respectively. In this case, the activation energy for the removal of the hydrating water molecules is less than the thermal energy. Silicotungstate is a known superchaotrope; i.e., it disrupts the water hydrogen-bond network around it (Assaf and Nau, 2018). This can facilitate the interaction between dopamine and silicotungstate; however, the relative importance of superchaotropicity decreases with increasing temperatures as the thermal energy of the water molecules increases.

In this work, we have used various complementary techniques to investigate the interactions between the polyoxometalate silicotungstic acid as a model for metal oxide nanoparticles and the biological signal substance dopamine in the solid state and in solution. Hydrogen bonding to the POM by the ammonium group and hydroxyl groups of dopamine was found. DLS and DOSY NMR both indicate that dopamine associates with silicotungstate even in solution, thus suggesting a strong interaction between the two molecules. This knowledge is of interest as nanoscale metal oxide nanoparticles used in in vivo medical applications may interact with the neurotransmitters, dopamine and norepinephrine, or other related metabolites found in the brain or the PMS. The temperature studies suggest a strengthened interaction between dopamine and silicotungstate at elevated temperatures, which could have biological implications in fever situations.

The current study, however, used a simple system consisting only of silicotungstate and dopamine in an acidic medium to promote crystallization. The silicotungstic POM serves as a molecular model system for dopamine interaction with natural silica and silicate particles, without strong catechol interaction. Real biological media (e.g., blood) are far more complex and may contain other biomolecules with potentially higher affinity for silicotungstate than dopamine. Future work should focus on the interactions between Keggin POMs and dopamine in physiologically relevant conditions and in the presence of other, competing biomolecules.